Dynamic face recognition device for intelligent construction site based on BIM model

Abstract

The invention relates to the technical field of face recognition, and discloses a BIM model-based dynamic face recognition device for an intelligent construction site, and the device comprises a sensing terminal which is integrated on a standard safety helmet and comprises an internal micro camera, an external wide-angle camera and a controller, the internal micro camera is fixed at the inner side of the brim of the safety helmet, and the external wide-angle camera is fixed at the outer side of the brim of the safety helmet; the camera is used for capturing stable eyebrow and forehead area images of a wearer in a non-intrusive mode and used for collecting a first biological characteristic image of an operator, and the external wide-angle camera is installed on the forehead of the safety helmet and used for collecting a first environment image of the view angle of the operator. A first environment image is collected through an external wide-angle camera to determine a space-time anchor point of an operator, a BIM real-time interaction engine inferes an expected operator according to the space-time anchor point, a first biological characteristic image of a local part such as unshielded eyebrows and eyes is collected through an internal micro camera, and the first biological characteristic image is compared with a preset biological characteristic model of the expected operator; and the stability and the accuracy of identity recognition are ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of face recognition, in particular to a dynamic face recognition device for smart construction site based on BIM model. BACKGROUND

[0002] The construction of a smart construction site puts forward higher requirements for the fine and automatic management of on-site personnel. The existing face recognition technology, as a key entrance, plays an important role in static scenes such as access control. The current face recognition application in the field of smart construction site is mainly fixed in several typical forms: fixed entrance management system, which deploys face recognition gates or access control at the entrances of construction site gates, dormitory buildings or key material warehouses. When personnel pass through, the system captures facial images and compares them with the background whitelist to realize attendance and basic regional access control. Shallow linkage with BIM model, the BIM model is used as a static database of personnel information. When the face recognition system confirms the identity of the personnel, it queries the BIM database for the personnel's job type, team and preset access rights to determine whether they can enter a specific area. Handheld terminal inspection application, management personnel use the App on the mobile phone or tablet computer to take pictures of on-site workers, manually or semi-automatically verify their identities, and associate the inspection results with the components or location points in the BIM model.

[0003] However, once the personnel enter the interior of the vast, dynamic and harsh construction site, the application has many defects: complex environmental factors such as strong light, backlight, shadow, rain, fog and dust seriously interfere with image acquisition quality and reduce recognition accuracy. Due to safety production requirements, workers must wear safety helmets, masks, protective glasses and other personal protective equipment, resulting in a large area of missing key facial features, and traditional recognition algorithms are invalid. The existing technology solves the "admission" problem, but cannot dynamically and accurately associate the identity of the personnel with their real-time location in the three-dimensional physical space represented by the BIM model and the construction tasks they undertake. The application of BIM model remains in the shallow static data query. SUMMARY

[0004] In view of the shortcomings of the prior art, the present application provides a dynamic face recognition device for smart construction site based on BIM model, which solves the bottleneck problem of identity recognition in complex environment of construction site.

[0005] To achieve the above purpose, the present application realizes the following technical scheme: a dynamic face recognition device for smart construction site based on BIM model, comprising: A perception terminal integrated on a standard safety helmet, comprising an inward-facing miniature camera, an outward-facing wide-angle camera, and a controller, the inward-facing miniature camera is fixed on the inner side of the brim of the safety helmet to capture stable images of the eyebrow, eye and forehead area of the wearer in a non-invasive manner, for collecting the first biological feature image of the worker, the outward-facing wide-angle camera is installed on the forehead of the safety helmet to collect the first environment image of the worker's visual angle, and the controller is in communication connection with the inward-facing miniature camera and the outward-facing wide-angle camera respectively; A BIM real-time interaction engine in communication connection with the controller and matching the first environment image collected by the outward-facing wide-angle camera with a preset BIM model to determine the space-time anchor point of the perception terminal in the BIM model; determining the expected worker associated with the space-time anchor point according to the space-time anchor point and the construction plan data contained in the BIM model; obtaining a preset biological feature model of the expected worker, and performing 1:1 comparison between the first biological feature image collected by the inward-facing miniature camera and the preset biological feature model to complete the identity recognition of the worker.

[0006] Preferably, the controller comprises a multi-source positioning and attitude sensing unit and a wireless communication module; wherein the multi-source positioning and attitude sensing unit integrates an inertial measurement unit to assist in sensing the head posture of the wearer and fuse other positioning signals as an auxiliary; the wireless communication module is used for data exchange with the BIM real-time interaction engine.

[0007] Preferably, the BIM real-time interaction engine comprises: a visual positioning and matching service module for performing matching between the first environment image and the BIM model to determine the space-time anchor point; a personnel information and feature database for storing the preset biological feature model; a collaborative verification and decision logic center for determining the expected worker according to the space-time anchor point and scheduling the 1:1 comparison.

[0008] Preferably, the first biological feature image is an image of a local area of the face of the worker that is not blocked when wearing personal protective equipment, and the preset biological feature model is a model previously established based on the image of the local area of the face of the worker.

[0009] Preferably, the BIM real-time interaction engine is specifically used for: processing the first environment image into an environment fingerprint; matching the environment fingerprint with a preset virtual environment fingerprint library generated according to the BIM model to determine the space-time anchor point.

[0010] Preferably, the BIM real-time interaction engine is further configured to: trigger peer cross verification when it is determined that the spatio-temporal anchors of the two aware terminals are in a proximity state; The peer cross verification comprises: instructing the outward-facing wide-angle camera of one of the aware terminals to capture a second environment image containing the wearer of the other aware terminal, and verifying the second environment image based on a preset biological feature model of the wearer of the other aware terminal.

[0011] Preferably, according to the result of the peer cross verification, a spatio-temporal trust chain is established for a plurality of aware terminals that have successfully verified each other, and the credibility of the identity recognition result is improved.

[0012] Preferably, after completing the identity recognition, the BIM real-time interaction engine is further configured to: determine a key work object corresponding to the current task of the expected worker according to the spatio-temporal anchor and construction plan data in the BIM model; feed the visual feature of the key work object as an attention focus instruction to the aware terminal in advance, so that the aware terminal judges whether the field of view of the worker is effectively focused on the key work object through the outward-facing wide-angle camera.

[0013] Preferably, the controller further comprises an edge computing unit configured to receive the attention focus instruction and analyze the image captured by the outward-facing wide-angle camera to determine whether the field of view of the worker is effectively focused on the key work object.

[0014] Preferably, the aware terminal further comprises a power supply which is detachably mounted on both sides of the safety helmet and electrically connected with the inward-facing miniature camera, the outward-facing wide-angle camera and the controller, respectively.

[0015] The application provides a dynamic face recognition device for a smart construction site based on a BIM model. The device has the following advantages: 1、The present application can reliably identify the operating personnel wearing a complete set of personal protective equipment, and skillfully avoids the industry problem of recognition failure caused by face covering. Its implementation mode is not to seek recovery or recognition of the face under the covering, but through a logical conversion: the external wide-angle camera of the device first collects the first environment image to determine the space-time anchor point of the operating personnel, and the BIM real-time interaction engine reasons out the expected operating personnel accordingly. This process converts an open "search" problem into a clear "verification" problem, that is, only by collecting the unblocked local first biometric feature image such as eyebrows and eyes through the internal miniature camera, and comparing it with the preset biometric feature model of the expected operating personnel can the 1:1 comparison be made. Since the face area required for verification remains exposed when wearing equipment, the stability and accuracy of identity recognition are ensured without violating safety procedures.

[0016] 2、The present application realizes continuous and dynamic tracking and management of all personnel inside the construction site, and expands the range of identity awareness from the isolated entrance "point" to the entire three-dimensional construction space "body". This is due to the continuous collection of first environment images by the external wide-angle camera of the sensing terminal, and the matching of the images with the BIM model by the back-end BIM real-time interaction engine, thereby generating an uninterrupted space-time anchor point track. This mechanism enables management personnel to accurately grasp the precise distribution and historical trajectory of each operating personnel in the three-dimensional space of the construction site, providing an unprecedented global dynamic view for resource scheduling, safety management and emergency response.

[0017] 3、The present application greatly deepens the application level of BIM model, making it change from a static three-dimensional information board to a "digital twin" engine that can actively perform cognition, reasoning and decision-making. When the BIM real-time interaction engine obtains a space-time anchor point, it not only passively displays the position, but also actively queries the associated construction plan data to reason out the expected operating personnel at that space-time. BIM is the core of logical reasoning, which provides context for events occurring on site and actively initiates the identity verification process. This deep integration enables BIM to truly participate in the intelligent construction site management system.

[0018] 4、The present application significantly enhances the reliability and anti-fraud capability of the entire system by introducing a collaborative verification mechanism between groups. When the system determines that the space-time anchor points of multiple sensing terminals are adjacent to each other, it will trigger a peer cross-verification program, i.e., the devices of the operating personnel will become temporary verification nodes for each other's identity. The device group that successfully completes the cross-verification will form a high-trust space-time trust chain. This decentralized verification network can effectively identify and isolate false information caused by single device failure, signal drift or malicious video fraud, establishing a solid "firewall" for the stable operation of the entire system.

[0019] 5、The application of identity management is improved from the confirmation of "personnel existence" to the supervision of "task execution". After confirming the identity and location of the worker, the BIM real-time interaction engine can further determine the key work object that the worker should focus on according to the construction plan and issue the attention focus instruction. The perception terminal analyzes the picture of the wide-angle camera to determine whether the worker's line of sight is focused on the key object. This makes the manager not only know "who is where", but also understand "whether he is doing the right thing" to a certain extent, which provides a new and fine data support for the quality control and process compliance management of the key process BRIEF DESCRIPTION OF DRAWINGS Figure 1 The device operation flowchart of the application; Figure 2 The perception terminal in the application is a top view stereographic diagram; Figure 3 The perception terminal in the application is a bottom view stereographic diagram.

[0020] 1, safety helmet; 2, internal miniature camera; 3, external wide-angle camera; 4, controller. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0022] Please refer to the drawings in the embodiments of the application Figure 1 - the drawings in the embodiments of the application Figure 3 The embodiment of the application provides a dynamic face recognition device for a smart construction site based on a BIM model, which comprises: The perception terminal is integrated on the standard safety helmet 1, and includes an inward miniature camera 2, an outward wide-angle camera 3 and a controller 4. The inward miniature camera 2 is fixed to the inner side of the brim of the safety helmet 1 to capture the stable eyebrow-eye-forehead area image of the wearer in a non-invasive manner, and is used to collect the first biological feature image of the worker. The first biological feature image is the image of the local area of the face of the worker that is not blocked when wearing personal protective equipment. The outward wide-angle camera 3 is installed on the forehead of the safety helmet 1 and is used to collect the first environment image of the worker's visual angle. The controller 4 is in communication connection with the inward miniature camera 2 and the outward wide-angle camera 3 respectively. The perception terminal further includes a power supply which is detachably installed on both sides of the safety helmet 1 and is in electrical connection with the inward miniature camera 2, the outward wide-angle camera 3 and the controller 4 respectively. The controller 4 includes a multi-source positioning and attitude sensing unit and a wireless communication module. The multi-source positioning and attitude sensing unit integrates an inertial measurement unit to assist in sensing the head posture of the wearer and fuse other positioning signals as an auxiliary. The wireless communication module is used for data exchange with the BIM real-time interaction engine. The controller 4 further includes an edge computing unit which is configured to receive an attention focus instruction and analyze the image collected by the outward wide-angle camera 3 to determine whether the field of view of the worker is effectively focused on the key work object.

[0023] The BIM real-time interaction engine is in communication connection with the controller 4 and matches the first environment image collected by the outward wide-angle camera 3 with the preset BIM model to determine the space-time anchor point of the perception terminal in the BIM model. According to the space-time anchor point and the construction plan data contained in the BIM model, the expected worker associated with the space-time anchor point is determined. The preset biological feature model of the expected worker is obtained, and the preset biological feature model is a model previously established based on the local area image of the face of the worker. The first biological feature image collected by the inward miniature camera 2 is compared with the preset biological feature model in a 1:1 manner to complete the identity recognition of the worker.

[0024] The BIM real-time interaction engine includes: a visual positioning and matching service module for performing matching of the first environment image and the BIM model to determine the space-time anchor point; a personnel information and feature database for storing the preset biological feature model; a collaborative verification and decision logic center for determining the expected worker according to the space-time anchor point and scheduling the 1:1 comparison.

[0025] The BIM real-time interaction engine is specifically used for processing the first environment image into an environment fingerprint, and matching the environment fingerprint with a virtual environment fingerprint library that is preset and generated according to the BIM model to determine the space-time anchor point.

[0026] When it is determined that the spatiotemporal anchors of two perception terminals are in a proximate state, peer cross-verification is triggered; the peer cross-verification includes: instructing the outward-facing wide-angle camera 3 of one of the perception terminals to capture a second environmental image containing the wearer of the other perception terminal, and verifying the second environmental image based on a preset biological feature model of the wearer of the other perception terminal. According to the results of the peer cross-verification, a spatiotemporal trust chain is established for multiple perception terminals that successfully verify each other, and the credibility of the identity recognition results is improved.

[0027] After the BIM real-time interaction engine completes the identity recognition, it is further configured to: According to the spatiotemporal anchor and the construction plan data in the BIM model, a key work object corresponding to the current task of the worker is determined; the visual features of the key work object are used as attention focus instructions and are fed forward to the perception terminal, so that the perception terminal can determine whether the worker's field of view is effectively focused on the key work object through the outward-facing wide-angle camera 3.

[0028] The above embodiment provides a dynamic face recognition device for a smart construction site based on a BIM model. The overall architecture of the device embodies a distributed system that works collaboratively between a front-end perception and a back-end cognition when implemented. The system is logically divided into two core entities: one is a perception terminal configured on a worker for on-site perception; the other is a BIM real-time interaction engine serving as the central decision-making hub of the system cognition. These two entities exchange data in real time and bidirectionally through a wireless communication network, and together form a complete dynamic recognition and verification closed loop.

[0029] In a preferred embodiment, the perception terminal has a physical form of a lightweight and low-power intelligent module that is integrated or fixed on a safety helmet 1 that must be worn by a construction worker. This deployment ensures that the device can move freely with the worker throughout the construction site and continuously acquire environmental and biological feature information of the worker in a first-person perspective. The combination of the perception terminal and the safety helmet 1, a standardized personal protective equipment, not only ensures the universal applicability of the device, but also makes its deployment and use in real work scenarios natural and non-intrusive, thereby providing a physical carrier for the dynamic tracking and recognition of workers throughout the construction site.

[0030] The BIM real-time interaction engine is usually deployed as a set of backend server systems. According to the specific project requirements and network conditions, the engine can be deployed on cloud servers with powerful computing and storage capabilities to support concurrent access and data processing for large-scale construction sites, or deployed as a localized server in the on-site computer room to obtain lower data transmission delay and stronger local data management capabilities. Regardless of the deployment method, the engine serves as the "digital brain" of the entire device, responsible for performing the most complex computing tasks, including processing and analyzing all data uploaded by the front-end sensing terminals, and deeply interacting with the massive BIM model, ultimately making identification, verification, and decisions.

[0031] The overall working mode of the device embodies a clear division of labor and cooperation. The sensing terminal located at the front end focuses on performing its core sensing and collecting tasks, just like the "eyes" and "senses" of the human body, responsible for capturing raw images and data. The BIM real-time interaction engine at the back end plays the role of "brain", which receives sensory information from the front end and uses its powerful cognitive and reasoning capabilities to transform it into a deep understanding of personnel identity, spatial location, and behavior state. The sensing terminal sends the collected data stream to the BIM real-time interaction engine through the 5G or Wi-Fi wireless network covering the construction site; the engine completes analysis, reasoning, and decision-making, and then returns necessary instructions or verification requests to the sensing terminal. This architectural design skillfully balances the relationship between the portability and endurance of the front-end device and the powerful data processing capabilities of the back-end system.

[0032] The sensing terminal is a highly integrated module that contains multiple units that work together. First, there is a two-way sensing camera system, which consists of an external camera (external wide-angle camera 3) and an internal camera (internal miniature camera 2). The external camera uses a wide-angle lens to obtain the widest possible first-person perspective of the worker, and the first environmental image it captures is the basis for subsequent visual positioning. The internal camera uses a miniature camera and is precisely installed on the inside of the safety helmet's brim, with its lens designed to stably capture the area above the worker's face, such as the eyebrow arch and forehead, even when wearing personal protective equipment. The first biometric image it captures is derived from this, providing a basis for subsequent local feature comparison.

[0033] The internal of the controller 4 in the perception terminal is also provided with an edge computing and positioning unit. The core of the unit is a low-power artificial intelligence (AI) processor, which is responsible for local preprocessing before data transmission. For example, it directly processes the raw video stream collected by the wide-angle camera 3 outside, and converts it into an environment fingerprint with small data volume and significant features by running a specific feature extraction algorithm, instead of transmitting large video files. At the same time, it can also execute the 1:1 comparison instructions issued by the BIM real-time interaction engine. The unit also integrates an inertial measurement unit (IMU) for real-time sensing of the head posture changes of the wearer. These posture data can be used as auxiliary information for visual positioning to improve the stability and accuracy of positioning.

[0034] To realize the connection with the backend, the perception terminal controller 4 also contains a wireless communication module. The module supports the 5G or high-density Wi-Fi network deployed on the construction site, and is responsible for establishing a low-latency, high-bandwidth data link between the perception terminal and the BIM real-time interaction engine, ensuring reliable upload of environment fingerprints and timely receipt of backend instructions.

[0035] As the cognitive decision-making center of the system, the BIM real-time interaction engine is internally composed of multiple software function modules that work together. Among them, the BIM model and the construction plan database are the knowledge cornerstone of the system. This database not only stores the detailed three-dimensional geometric model of the building, but more importantly, it also integrates 4D construction plan data, which accurately binds specific construction tasks, execution time, responsible person, etc. with specific components or spatial positions in the BIM model.

[0036] The visual positioning and matching service module is the core of the engine to realize spatial cognition. During the system initialization phase, this service will generate a virtual environment fingerprint library covering the entire construction site area by automatically rendering the BIM model, containing thousands of virtual environment fingerprints with accurate three-dimensional coordinates and orientation labels. During system operation, after receiving the real-time environment fingerprint uploaded by the front end, the service will use efficient image retrieval technology to quickly match in the virtual fingerprint library. Once a match is found, the current space-time anchor point of the perception terminal can be immediately determined.

[0037] The personnel information and biometric database is responsible for the management of personnel data. It stores information of all authorized personnel entering the construction site, and each person is associated with a pre-set biometric model established based on the image of a local area of their face, which has been encrypted. This model serves as a reference for subsequent 1:1 identity verification.

[0038] Finally, the collaborative verification and decision-making logic center is the "commander" of the entire engine. This module is responsible for orchestrating and scheduling the work of all other modules. It receives the spatio-temporal anchor points determined by the visual positioning service, then queries the BIM database to infer the expected personnel, and then retrieves the corresponding preset biometric model from the personnel database and issues a 1:1 verification instruction. In addition, all high-level intelligent decisions such as triggering peer cross-verification, establishing and managing spatio-temporal trust chains, and issuing attention focus instructions are all completed by this logic center. It organically links the data and functions of various modules to form the unique, layer-by-layer cognitive and verification process of the invention.

[0039] When the worker first registers and receives the perception terminal, the system will perform an identity initialization and local biometric modeling process. In this process, the worker needs to face the inward-facing miniature camera 2 of the perception terminal, which will collect multiple frames and multiple angle first biometric images of the area above his face. The reason for choosing this area is that it can still maintain a stable, unobstructed exposure state after the worker wears personal protective equipment such as safety helmets and masks according to regulations. After receiving these images, the BIM real-time interactive engine processes them into a compact, robust, and unique identity-representing preset biometric model through feature extraction algorithms. This model is then encrypted and bound with the worker's identity information (such as name, job type, permissions, etc.) in the BIM system and stored in the personnel information and biometric database, providing a reference for subsequent 1:1 verification.

[0040] When the worker wears the perception terminal and moves around the construction site, the system enters a continuous, environment fingerprint-based spatio-temporal anchor point positioning process. The outward-facing wide-angle camera 3 of the perception terminal continuously collects first-person perspective first environment images. The edge computing unit built into the controller 4 does not directly transmit raw video, but runs an efficient feature point extraction algorithm to extract key points and descriptors that remain stable under changes in lighting and perspective from the image, and then encodes them into a data-minimized environment fingerprint before uploading it through the wireless module. The visual positioning and matching service module of the BIM real-time interactive engine will immediately perform high-speed search and matching in the pre-constructed BIM virtual environment fingerprint library upon receiving this environment fingerprint. Once the best matching virtual fingerprint is found, the precise three-dimensional coordinates and orientation information attached to it are determined as the spatio-temporal anchor point of the perception terminal at the current time.

[0041] Subsequently, the system seamlessly connects to the 1:1 identity verification process based on BIM active reasoning. This process is the core of the invention to solve the problem of occlusion recognition. At the moment when the spatiotemporal anchor point is established, the collaborative verification and decision logic center of the BIM real-time interaction engine will immediately use this spatiotemporal information as an index to query the BIM model and the construction plan database. For example, the query result shows that at the current time, the construction task corresponding to this three-dimensional coordinate point is "pipe welding in A area of the second floor", which should be executed by welder "Zhang San" according to the plan. The system makes active reasoning based on this and determines "Zhang San" as the expected worker at this position. Then, the engine retrieves the preset biological feature model of "Zhang San" from the database and issues it as a 1:1 verification instruction to the sensing terminal at this position. After receiving the instruction, the terminal instantaneously collects the current first biological feature image of the wearer through the internal miniature camera 2, and compares it with the received model locally. The comparison result (yes / no) is returned to the engine, thereby completing the identity recognition with high efficiency and accuracy.

[0042] To further improve the overall credibility of the system, the invention also includes a peer cross-verification process based on group perception. When the BIM real-time interaction engine determines that multiple sensing terminals form an aggregation in the physical space (for example, a team works on the same work platform) through spatiotemporal anchor point data, it will actively trigger this process. The engine will instruct the first sensing terminal to use its external wide-angle camera 3 to find the wearer of the second sensing terminal in its field of view, and use the preset biological feature model of the wearer of the second sensing terminal obtained from the engine to perform a temporary identity verification. This process is carried out within the group. When multiple terminals in an area successfully verify each other's identities, the engine will establish a high-credibility spatiotemporal trust chain for this temporary group. The credibility of the positioning and identity of any individual who cannot be confirmed by the neighboring verified nodes will be automatically reduced by the system, and attention will be triggered.

[0043] Finally, after the identity and location of the worker are both confirmed with high confidence, the system can extend the verification to the task execution level, i.e., initiate the attention focus verification process based on BIM tasks. If the current spatio-temporal anchor point corresponds to a key procedure in the BIM construction plan, the BIM real-time interaction engine extracts the core operation object of the procedure from the model, such as a specific high-voltage switch or a prefabricated component that needs to be installed precisely, and defines its three-dimensional model or visual features as the key operation object. Subsequently, the engine feeds the visual features of this object as an attention focus instruction to the worker's perception terminal. The edge computing unit on the terminal analyzes the video stream of the outward-facing wide-angle camera 3 to determine whether the wearer's field of view center has been effectively time-stayed and focused on the designated key operation object. This completes the final closed-loop verification of "the right person, at the right place, focusing on the right thing".

[0044] To more intuitively demonstrate the collaborative working mode of each technical link in the device of the present application, a typical application scenario is described below. The scenario describes the complete process of a worker from entering the site to performing a key procedure.

[0045] Suppose an electrician whose identity information and corresponding preset biological feature model has been pre-registered in the system. When the worker wears a safety helmet that has been enabled and integrated with a perception terminal to enter the construction site, the device of the present application starts to work. Its outward-facing wide-angle camera 3 starts to automatically collect the first environmental image of the worker's first-person perspective, and the edge computing unit in the terminal processes it into a series of continuous environmental fingerprints in real time and uploads them to the BIM real-time interaction engine in the back end through a wireless network.

[0046] Upon receiving the first environmental fingerprint, the engine immediately performs matching in the vast BIM virtual environmental fingerprint library and quickly determines the initial spatio-temporal anchor point of the worker, such as inside the main entrance of the site. Then, the decision logic center of the engine queries the BIM construction plan database according to the spatio-temporal anchor point and the current time, infers that the entry behavior corresponds to the task of "electrician team entry" in the plan, and determines the expected worker identity of the worker. A 1:1 verification instruction containing the preset biological feature model of the worker is immediately issued to the worker's perception terminal.

[0047] After receiving the instruction, the perception terminal collects the first biometric image of the worker's eyebrow and eye area through the internal miniature camera 2, and completes the comparison with the issued model locally. The consistent result is returned, and the identity of the worker is confirmed for the first time. Thereafter, as the worker walks inside the construction site to his work point on the fifth floor of the building, the above-mentioned "environment positioning-identity reasoning-verification" process continues to circulate at a high frequency, forming a real-time three-dimensional moving track inside the BIM model.

[0048] When the worker arrives at the fifth floor work area, his spatiotemporal anchor point shows that he is in a neighboring state with another worker who is installing a pipeline nearby. After recognizing this spatial relationship, the decision logic center of the BIM real-time interaction engine initiates a cross-verification program. The engine sends the preset biometric model of the other worker to the perception terminals of the two workers respectively. The perception terminal of the electrician worker captures the image of the pipeline worker through the external wide-angle camera 3 and completes the verification, and vice versa. After the two successful verification results are reported, the system establishes a temporary spatiotemporal trust chain for them, further consolidating the credibility of their respective position and identity data.

[0049] Finally, the electrician worker arrives at the designated distribution box in the construction plan, and his spatiotemporal anchor point completely coincides with the coordinates of the distribution box in the BIM model, and his identity is also re-confirmed. At this time, the verification level of the BIM real-time interaction engine is promoted from identity confirmation to task confirmation. According to the construction plan, the engine determines that the core operation object of its current task is a specific circuit breaker in the distribution box, which is defined as the key operation object.

[0050] The engine feeds the visual features of the circuit breaker, such as its shape, color and label text, to the perception terminal of the electrician as an attention focus instruction. When the worker opens the distribution box door and focuses his gaze on finding and operating the specific circuit breaker, the edge computing unit of his perception terminal analyzes the video stream of the external wide-angle camera 3 to determine whether the center of the field of view is effectively stopped on the specified key operation object. Once confirmed, the entire verification process forms a complete closed loop: from "the right person" to "the right place", and then to "focus on the right thing", each link is confirmed by data support.

[0051] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic facial recognition device for smart construction sites based on BIM models, characterized in that, include: The sensing terminal, which is integrated on a standard safety helmet (1), includes an internal miniature camera (2), an external wide-angle camera (3), and a controller (4). The internal miniature camera (2) is fixed to the inside of the brim of the safety helmet (1) and captures stable images of the wearer's eyebrows, eyes, and forehead area in a non-invasive manner to collect the first biometric image of the worker. The external wide-angle camera (3) is installed on the forehead of the safety helmet (1) and is used to collect the first environmental image from the worker's perspective. The controller (4) is communicatively connected to the internal miniature camera (2) and the external wide-angle camera (3). The BIM real-time interactive engine communicates with the controller (4) and matches the first environmental image collected by the external wide-angle camera (3) with the preset BIM model to determine the spatiotemporal anchor point of the sensing terminal in the BIM model; based on the spatiotemporal anchor point and the construction plan data contained in the BIM model, it determines the expected workers associated with the spatiotemporal anchor point; it obtains the preset biometric model of the expected workers and compares the first biometric image collected by the internal miniature camera (2) with the preset biometric model 1:1 to complete the identification of the workers.

2. The dynamic face recognition device for smart construction sites based on BIM models according to claim 1, characterized in that, The controller (4) includes a multi-source positioning and attitude sensing unit and a wireless communication module; wherein, the multi-source positioning and attitude sensing unit integrates an inertial measurement unit to assist in sensing the wearer's head posture and integrates other positioning signals as an aid; the wireless communication module is used to exchange data with the BIM real-time interactive engine.

3. The dynamic face recognition device for smart construction sites based on BIM models according to claim 1, characterized in that, The BIM real-time interactive engine includes: The visual positioning and matching service module is used to perform matching between the first environmental image and the BIM model to determine spatiotemporal anchor points; A personnel information and feature database is used to store the preset biometric model; The collaborative verification and decision-making logic center is used to determine the expected operators and schedule the 1:1 comparison based on the spatiotemporal anchor points.

4. A dynamic face recognition device for smart construction sites based on BIM models according to claim 1, characterized in that, The first biometric image is a partial image of the worker's face that is not obscured when the worker is wearing personal protective equipment, and the preset biometric model is a model pre-established based on this partial image of the worker's face.

5. A dynamic face recognition device for smart construction sites based on BIM models according to claim 1, characterized in that, The BIM real-time interactive engine is specifically used for: The first environmental image is processed into an environmental fingerprint; The environmental fingerprint is matched with a preset virtual environmental fingerprint library generated based on the BIM model rendering to determine the spatiotemporal anchor point.

6. A dynamic face recognition device for smart construction sites based on BIM models according to claim 1, characterized in that, The BIM real-time interaction engine is also configured to: When it is determined that the spatiotemporal anchor points of two sensing terminals are in a proximity state, peer-to-peer cross-validation is triggered. The peer-to-peer cross-validation includes: instructing one of the sensing terminals to use its external wide-angle camera (3) to capture a second environmental image containing the wearer of the other sensing terminal, and verifying the second environmental image based on the preset biometric model of the wearer of the other sensing terminal.

7. A dynamic face recognition device for smart construction sites based on BIM models according to claim 6, characterized in that, Based on the results of the peer-to-peer cross-validation, a spatiotemporal trust chain is established for multiple sensing terminals that have successfully verified each other, thereby improving the credibility of their identity recognition results.

8. A dynamic face recognition device for smart construction sites based on BIM models according to claim 2, characterized in that, After completing identity recognition, the BIM real-time interactive engine is further configured as follows: Based on the spatiotemporal anchor points and the construction plan data in the BIM model, the key work objects corresponding to the current tasks of the expected workers are determined; The visual features of the key work object are used as attention focus instructions and fed forward to the sensing terminal so that the sensing terminal can determine whether the field of view of the operator is effectively focused on the key work object through the external wide-angle camera (3).

9. A dynamic face recognition device for smart construction sites based on BIM models according to claim 8, characterized in that, The controller (4) further includes an edge computing unit configured to receive the attention focus instruction and analyze the image captured by the external wide-angle camera (3) to determine whether the operator's field of view is effectively focused on the key work object.

10. A dynamic face recognition device for smart construction sites based on BIM models according to claim 1, characterized in that, The sensing terminal also includes a power supply, which is detachably installed on both sides of the safety helmet (1) and electrically connected to the internal miniature camera (2), the external wide-angle camera (3), and the controller (4).

Images